Elile AI revolutionizes energy management with an AI-first platform that ensures real-time adaptability and interoperability across energy assets. They enhance efficiency and reliability by predicting and preventing disruptions, aiming for a sustainable, intelligent energy ecosystem that integrates the grid with various energy systems.
How did your previous experience shape your vision for Elile?
Muhammad M Mehdi: I Have been part of the AI journey for over 20 years—hard to imagine given the recent hype, but my PhD was in AI, specifically in bin packing, which is about resource allocation under extreme uncertainty.
While completing my PhD, I built a company and sold it to a larger company called Intralinks, which then had an IPO. Following that, I joined a few friends from MIT to start building a wearable startup – a watch that could measure your blood pressure. We got funding from top VCs in United States.
Before 2010, I had three startups, all focused on Artificial Intelligence.
Afterwards, I joined Microsoft as one of the early engineers in what later became the cloud team. Back then, it was known as the Server and Tools group, led by a young vice president named Satya Nadella. In that group, we grew the cloud infrastructure from scratch.
Microsoft had no prior experience in building infrastructure; they only shipped DVDs of Office 365 or Windows. We had to learn how to build a global public utility. We assembled a global team from Europe, Britain, Germany, Spain, and Latin America, creating a network of 175,000 miles of cable—enough to go to the moon and back – four times. We also deployed over 20,000 megawatts of data centers worldwide.
Our strategic decisions, from deployment to partnerships, were driven by AI. My team managed extensive models and simulations. Azure’s CTO, Mark Russinovich, wrote a series of four blogs on what makes Microsoft the most reliable cloud in the world, with two of those blogs focused on the services my team built.
I have a strong background in infrastructure. Before leaving Azure for Microsoft Research, I designed the infrastructure for the nonprofit OpenAI. I joined Microsoft Research to lead the Agentic AI group, focusing on autonomous AI for industrial applications, including drone simulations and control systems. Later in my career, I worked on industrial applications of AI.
What brought you to this region, and why did you decide to establish your company in the UAE?
Muhammad M Mehdi: When I joined Microsoft, we had the opportunity to start from scratch with no existing cloud infrastructure. Building something from the ground up was an incredible rush and provided immense satisfaction. I wanted to replicate that experience, and the Middle East is the ideal region for that.
First, we wanted to apply the principles we used in the cloud industry to the energy sector. Second, we are eager to expand into Oman, Saudi and then go global. In the cloud, we started with basic technology and expanded it globally, making it interoperable and open source so that any vendor or operator could use it.
Now, we aim to bring the same cloud mindset to energy. The cloud grew exponentially, with deployments doubling every 14 months from 2012 to now. Our vision at Elile AI is to replicate this rapid growth in the energy sector.
Our broader thesis is “energy abundance.” We believe we are entering a new era where people will have so much energy that it will be liberating. This abundance will enable the development of better societies, communities, and more productive AI models. To achieve decentralized or sovereign AI in every region, access to ample energy is essential.
How has your experience in global media planning at Amazon influenced your approach to market strategy and customer engagement, especially in the context of promoting advanced energy solutions?
Muhammad M Mehdi: Amazon is renowned for its customer-centric approach, focusing on understanding customer needs and working backwards from those problems. At Amazon, I was head of media planning and AI tech for Amazon AdTech, where I saw the revenue for Amazon AdTech double and led the global expansion of media planning.
This experience taught me the importance of not only building great technology but also ensuring it has the right product-market fit and effective distribution channels. Without these elements, even the best technology won’t achieve the desired impact.
Can you discuss the impact of AI on global sustainability efforts?
Muhammad M Mehdi: We take the risks associated with AI very seriously, prioritizing AI security. Our solutions are designed with a strong emphasis on security to prevent AI models from becoming threat vectors or sources of intrusion. We follow the best security practices throughout the AI lifecycle—from training and deployment to operation—using AI itself to enhance security.
Addressing AI’s challenges involves a responsible approach and leveraging expertise from those who have managed global infrastructure. Involving a broad community is crucial to overcoming these challenges.
On the benefits side, AI helps us grow the energy sector exponentially while addressing risks like unreliable energy, outages, and efficiency gaps that drive up costs. One of the reasons we haven’t seen more widespread adoption of clean tech energy, hydrogen, and carbon-neutral options is their reliability compared to traditional methods. By using AI and the technology We are developing, we aim to make these options more reliable than traditional methods. Increased reliability boosts customer confidence and adoption, while also reducing costs.
What do you see as the next big breakthrough in AI technology?
Muhammad M Mehdi: The next big breakthrough in AI is likely to be more domain-specific models. We will see AI systems that specialize in understanding cultural, regional, and industrial contexts. Additionally, models will increasingly feature advanced decision-making capabilities with deeper decision trees, allowing them to reason through options and trade-offs more like human experts. This area is known as Neurosymbolic AI.
Neurosymbolic AI, along with advancements in Generative AI, will also see multi-agentic AI take off. This approach helps us understand the root causes behind AI decisions, reducing the risk of AI hallucination and improving our grasp of cause-and-effect relationships.
Moreover, we will witness significant growth in data access through enhanced data lakes and data lakehouse architectures. These technologies will enable the integration of vast amounts of data streams, offering new insights that were previously unattainable.
Elile is truly at the intersection of AI and clean tech. We are building a framework for using digital twins to alert, diagnose and optimize strategic and operational decisions— in this we are ahead of GE and Siemens. While we are also auto tuning the GE vernova digital twins — still a manual process everywhere else.
Together, these advancements will make AI more reliable and applicable in industrial settings.
How does this relate to customer or user protection, especially considering that AI requires a lot of data, which is often processed unconsciously? What is your position on this issue?
Muhammad M Mehdi: Data Privacy specially for organizations and sovereign states is critical. We believe that data ownership should remain with the organization or sovereign state that generated it. One approach we support is federated learning, a concept that originated at Google and which I advocated for while at Amazon.
Federated learning ensures that data remains with its owner. Instead of transferring raw data, only features derived from the data are shared. These features are then used to train AI models in a feature store, without accessing the underlying data itself. This approach gives individuals and organizations control over how their data is used.
With federated learning, you have continuous visibility into your data’s usage and can retract it if desired. This method empowers users to manage their data actively and make informed decisions about its use. By adopting such approaches, we aim to enhance data privacy and ensure that data sharing occurs with full transparency and mutual benefit, creating new business value while respecting privacy.
What sets your company apart in the market?
Muhammad M Mehdi: What differentiates us is our mindset. In the energy industry, as in many sectors, we often encounter isolated “walled gardens” with proprietary software and models that aren’t interoperable. While this approach has worked for a few Western companies, it has limited options and creates lock-in scenarios for the rest of the world.
We are bringing a cloud-inspired approach to the energy sector. In the cloud industry, I was involved in creating a framework where different clouds could interact seamlessly, enabling multi-cloud environments. This interoperability allows systems to deploy and onboard freely between different platforms.
We aim to apply this same principle to energy. Our platform will act almost invisibly but will support seamless integration with any new hardware or software. Think of it like an app store—where any new app can be downloaded and used easily.
In practice, this means we are bringing this open, interoperable model to various energy sectors, including solar, wind, water desalination, and gas turbines. For example, we are collaborating with GE Vernova to enhance their asset performance management (APM) platform with advanced analytics. Our AI will not be limited to GE but compatible with other vendors and OEM manufacturers.
This means that regardless of whether you have a GE or Siemens, or equipment from other manufacturers, our AI engine will integrate, support and optimize it. It will normalize data, provide insights, monitor performance, alert on issues, build digital twins, and optimize operations. This approach ensures greater flexibility, efficiency, and value across different energy systems.
Check our previous exclusive interview with Said Al Shanfari, CEO of Oman Convention and Exhibition Centre (OCEC).
